Total cost of ownership (TCO) is meant to show what a decision costs over its life, not only at purchase. Most TCO models are built from invoices: purchase price, energy, maintenance, spare parts, freight. Where an invoice exists, they are accurate.
Much of what a decision changes never appears on an invoice: a worker’s detour around a blocked aisle, a pallet rebuilt at the dock, a case that a machine hands back to a person. These costs sit in labor budgets and overhead, spread over several owners, and in most companies we know nobody adds them up. So the option that looks cheaper on paper is sometimes the more expensive one to run.
A process-oriented TCO starts from the process. It follows the work step by step, measures each step in physical units, values it at local cost rates and assigns it to the budget where it is booked. Only then do we look at the invoices.
Measure first, then value
We took the habit from automotive logistics, where a line is measured before anyone prices it. Our teams record time per process step, quantities, waiting, rework and damage, and they ask the people who do the work what goes wrong on a bad day. The cost data are added afterward. A process step without a cost is usually a cost that nobody books.
Keeping the physical effect apart from its cost rate is what makes the method portable. Seconds per unit or units per truck carry over from one site to the next. Wages, freight rates and prices do not, so every site is valued at its own rates.
Every effect takes the same path
Each effect passes through the same six fields before it enters the result.
- Process stageWhere in the process the effect arises
- Physical effectIn units: seconds per piece, pieces per hour, pallets per truck
- Cost rateThe local rate for labor, freight, material or space
- Budget and ownerWho books the cost, and who has to act for it to change
- Kind of effectCash, labor time, or a cost shifted to another party
- TimingOne-time or recurring
The fifth field is the one most business cases skip. Labor time becomes money only when the labor standard is updated and the hour budget falls, and that takes a decision by whoever owns the standard. A shifted cost saves nothing for the chain as a whole, though it tells you who has to agree.
Rules for the result
Two rules carry most of the weight. Only measured effects count: a gain that we saw but could not separate from other changes stays out, however plausible it looks. And losses are booked as carefully as gains, including a revenue that disappears with the old solution.
When a supplier or a vendor reports extra effort, we walk through the steps with them. Often the effort is real, sometimes it is habit. The one-time cost of a change is planned separately and does not enter the recurring result.
A range instead of one number
A single result hides its own uncertainty. We model what drives each cost and give every driver a distribution that fits what we know about it.
- Time measured on site
- Narrow spread around the measurement
- Market price
- The spread it showed over the past years
- Estimate
- Lowest, most likely and highest value
- Drivers that move together, such as wages along the chain
- Drawn together
A simulation draws many thousands of combinations. We report a conservative case, the median and an optimistic case, and the recommendation rests on the conservative one. The simulation also shows which driver makes the range wide. In our studies it was usually a market price; the measured process effects moved the result surprisingly little.
Some effects resist a number, such as how a team copes with a cramped backroom. We describe them next to the result and leave them out of it.
Three applications
The method is the same in every industry. What changes is where the costs hide.
Retail
One-way boxes or reusable containers for fresh produce
- Usual comparison
- Box price against the fee per container trip
- What the process adds
- Erecting and closing at the packer, pallet stability in the cold chain, rebuilt pallets and cleanup in the DC, breaking down and baling in the store, lost bale revenue, weight in freight
- Booked in
- Supplier, carrier, DC, store labor, waste contract, category management
Distribution center
Automated picking in place of manual picking
- Usual comparison
- Investment against picking hours saved
- What the process adds
- Items the system cannot handle, decanting into totes, inbound cases that stop the line, availability and maintenance, ramp-up, peaks
- Booked in
- Capital budget, DC operations, maintenance, IT, purchasing
Assembly
A robot cell in place of a manual station
- Usual comparison
- Cell and integration against operator hours
- What the process adds
- Part presentation, variants and changeovers, part tolerances, line balance, programming skills
- Booked in
- Capital budget, production, intralogistics, quality, engineering
Retail: one-way or reusable packaging
This is the case we know best, and the one where the gap between invoice and process is widest. Retailers usually compare a corrugated box with a reusable plastic container on one line, the price of the box against the fee for a container trip. Procurement owns both numbers.
What the box costs after purchase sits elsewhere. At the packer, staff erect and close it and secure the pallet. In the cold chain, thin board softens and stacks lose stability, so the DC rebuilds pallets and cleans up. In the store, someone breaks the box down, carries it to the baler and keeps the bale until it is hauled away. The reusable container has costs outside its fee too, above all its weight in the freight bill.
The process view also shows who books what. Part of any saving arises at the suppliers and reaches the retailer only through its purchase prices. Inside the retailer, the store can save hours while another unit loses the revenue from the bales. That decides who has to agree before the switch pays.
Distribution center: automating the picking
A business case for automated picking usually sets the investment against the picking hours it saves. What it tends to leave out is the work that grows around the system.
Goods often have to be decanted from supplier cases into totes first. Items that do not fit the system go to a manual area, which needs its own staff and space. Inbound cases that are out of dimension, soft or unstable stop the conveyors, so packaging rules for suppliers and checks at receiving become part of the cost. Then there is the ramp-up before the system reaches its rated output, and the peaks above its capacity that people still cover.
Decanting and exceptions eat into the hours saved. How much they eat is the number that decides the case, and it can be measured in the current operation before anything is bought.
Assembly: a robot cell on the line
Here the usual comparison is the cell and its integration against operator hours. The process view starts with the part: a robot needs it in a defined position, so the work of presenting it moves to intralogistics or to the supplier, as kitting, trays or feeders. A worker compensates part tolerances and variant changes by hand without anyone noticing. A cell stops. And once the station is faster, the bottleneck may simply move next door.
The cell can also raise quality where manual work varies, in screwing or bonding for example. That gain lands in the quality budget, the extra kitting in logistics.
What a decision needs
Most of the data already exists, spread across purchasing, logistics, production or store labor, maintenance and finance. Some has to be measured on site: times per step, exceptions, rework, and damage by cause.
The result is a cost and benefit breakdown by category, a range, and a decision matrix that weighs the options from the view of finance, operations and the unit that has to live with the outcome. Where the data for an effect does not exist yet, the matrix says so rather than filling the gap with an assumption.